By Topic

Performance evaluation studies of client-server models using SPEC Web99 benchmarks

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

4 Author(s)
Godbole, A. ; Electr. & Comput. Eng. Dept., Dayton Univ., OH, USA ; Seung-yun Kim ; Guzman, R. ; Smari, W.W.

The World Wide Web (WWW) has experienced tremendous growth over the past few years. This increase in use makes server load heavier and leads to more system-bound queries due to newly sought emerging Web techniques, such as information discovery and reuse, mining and fusion. There is an obvious interest in assessing the consequences of this growth and understanding the system's behavior under varying conditions and workloads. The overall performance of the WWW is affected by various criteria in terms of clients, servers, and the network. Measuring performance of a computer system is a complex, difficult and important task. Time and rates are usually the basic measures of system performance. Having more realistic data about system performance will be necessary in the design and development of integrated information systems and systems for reuse. The main objective of this work is to measure the performance of various Web-based computer systems and Web servers using the SPEC Web99 benchmarking toolset. In the experiments that we carried out, the server is set up using freely available Apache server software run on a Sun Sparc-Ultra 80 machine, and six Windows-based clients that are used to load the server using the benchmark suite. First, the effects of varying resources, such as memory sizes and processors, on server's performance are measured and analyzed. Then, we tested and evaluated the effects of six different clients and workload parameters on the server configuration. Amongst the six experiments that were performed, two were set up in order to test the client-side. Some of the parameters varied were memory sizes, number of processors, the workload, and the benchmark parameters. Some interesting results were obtained while others were as expected. The results indicate that good improvements can be achieved if workload and system resources are carefully chosen to optimize metrics of interest.

Published in:

Information Reuse and Integration, 2003. IRI 2003. IEEE International Conference on

Date of Conference:

27-29 Oct. 2003